eWeek Security News

1. OpenAI Debuts GPT-5.4-Cyber, a Locked-Down AI Model for Cyber Defense

   OpenAI launches GPT-5.4-Cyber with selective access for verified defenders, adding a cyber-focused model and tighter controls for high-risk use.

   The post OpenAI Debuts GPT-5.4-Cyber, a Locked-Down AI Model for Cyber Defense appeared first on eWEEK.

2. Anthropic Briefed Trump Administration on Mythos Cyber Capabilities

   Anthropic briefed senior Trump administration officials on its new Mythos model before giving outside organizations access, highlighting how seriously the company views the system’s cybersecurity implications. The model is not being released publicly, and Anthropic has positioned it as powerful enough to provide defensive value while also carrying clear risks of misuse. Mythos is at […]

   The post Anthropic Briefed Trump Administration on Mythos Cyber Capabilities appeared first on eWEEK.

3. AI Is Exposing a Skills Gap in Cybersecurity Hiring

   AI is transforming cybersecurity jobs by automating routine tasks and shifting roles toward decision-making, analysis, and advanced security skills.

   The post AI Is Exposing a Skills Gap in Cybersecurity Hiring appeared first on eWEEK.

4. Project Glasswing: Anthropic Unites Apple, Google, Microsoft on AI Cybersecurity

   Anthropic’s Project Glasswing unites major tech rivals to use Claude Mythos Preview to find and fix critical software vulnerabilities before attackers do.

   The post Project Glasswing: Anthropic Unites Apple, Google, Microsoft on AI Cybersecurity appeared first on eWEEK.

5. Microsoft Unveils $10B Plan to Scale AI, Cloud, and Cybersecurity in Japan

   Microsoft will invest $10 billion in Japan through 2029 to expand AI infrastructure, deepen cybersecurity ties, and train workers for cloud and AI roles.

   The post Microsoft Unveils $10B Plan to Scale AI, Cloud, and Cybersecurity in Japan appeared first on eWEEK.

6. Leaked Anthropic AI Model Raises ‘Unprecedented’ Security Concerns

   Leaked documents reveal Anthropic’s next AI model, Mythos, with major gains in coding and cybersecurity, raising new questions about AI risks.

   The post Leaked Anthropic AI Model Raises ‘Unprecedented’ Security Concerns appeared first on eWEEK.

7. This Week in AI: Meta Incident Highlights a Loss of Control

   Meta flags a Sev 1 incident after an AI agent posts data on its own, raising urgent questions about control, security, and who governs AI systems.

   The post This Week in AI: Meta Incident Highlights a Loss of Control appeared first on eWEEK.

8. Sam Altman-Backed Startup Aims to Verify Humans Behind AI Transactions

   World introduces AgentKit to verify humans behind AI shopping agents, aiming to reduce fraud and improve trust in automated commerce.

   The post Sam Altman-Backed Startup Aims to Verify Humans Behind AI Transactions appeared first on eWEEK.

9. Meta Boosts Creator Protection on Facebook

   Meta adds stronger anti-impersonation tools on Facebook while rewarding original creators and pushing copycats and reposted content further down.

   The post Meta Boosts Creator Protection on Facebook appeared first on eWEEK.

10. Russian-Linked Phishing Campaign Targets Signal, WhatsApp Users

    Dutch intelligence says Russian state hackers are targeting Signal and WhatsApp users through phishing, fake support messages, and device-linking scams.

    The post Russian-Linked Phishing Campaign Targets Signal, WhatsApp Users appeared first on eWEEK.